The availability of high-dimensional single-nucleotide (SNP) data for large samples of individuals with and without disease presents unprecedented opportunities for genetic epidemiologists, but also many statistical challenges. Our ability to bring these opportunities to fruition depends on the development and application of sound and innovative statistical methods. Areas of particularly great need include methods for dealing with multiple hypothesis-testing problems (multiple comparison issues), and methods for combining data from multiple sources to achieve sharper inferences (data synthesis issues). The goals of this research are to develop new or improved methods for addressing these issues, and to apply them to data on cancers of the breast, ovary and prostate. To accomplish these goals, the investigators will build on more than 15 years of previous work. Specifically, in 1988 the National Cancer Institute (NCI) awarded an Outstanding Investigator Grant (OIG) to the Principle Investigator for the development and application of new and improved statistical methods for use in epidemiological research. In 1995 this grant was renewed until NCI terminated the program in 2001. The research was then funded by R01 CA94069 for the period January 2002 to December 2006. This competing renewal application requests funds to continue this work and to apply it to new areas of emerging importance. The specific aims are fourfold: 1) to evaluate new and existing methods for controlling confounding and tail count variability in case-control genome-wide association (GWA) studies; 2) to evaluate risks associated with missense mutations of unknown significance in genes of established disease relevance; 3) to combine multiple independent sources of data in assessing the joint carcinogenic effects of groups of genes; and 4) to integrate data on tumor and patient characteristics for improved assessment of phenotype-specific etiology. We will evaluate the new methods by simulations, and we will illustrate them by application to existing data on cancers of the breast, ovary and prostate, and emerging data from GWA studies of breast and prostate cancer. The existing data have been collected either by the investigators and the consultants to this project, or as part of three collaborations in which the investigators participate: the Breast Cancer Family Registry (Breast-CFR), the Ovarian Cancer Association Consortium (OCAC), and the International Consortium for Prostate Cancer Genetics (ICPCG). In summary, we need sound, reliable methods for analyzing the vast amounts of emerging genetic data from individuals with and without a given type of cancer. The goals of this research are to develop such methods, and thereby to help generate new knowledge useful for cancer prevention and treatment. [unreadable] [unreadable] [unreadable] [unreadable] [unreadable] [unreadable] [unreadable] [unreadable] [unreadable] [unreadable] [unreadable] [unreadable] [unreadable]